Pentagon AI Security Push, Nuclear Waste Challenges, and China’s AI Tensions

The Pentagon is preparing to establish secure environments for generative AI companies to train military-specific versions of their models on classified data, MIT Technology Review has learned. This represents a major new development that brings AI firms closer to classified intelligence than ever before. The classified training would embed sensitive information—such as surveillance reports or battlefield assessments—directly into AI models themselves. Current systems like Anthropic’s Claude are already used to answer questions in classified settings, including analyzing targets in Iran, but allowing them to learn from classified data presents unique security risks.

Reported by: Julian Reed

Consumer Electronics Expert